132 ON THE EFFECT OF THE INTERNAL FRICTION OF FLUIDS and t~tf for t in (18*7), and integrating with respect to t'. We thus get To apply this result to the case of an oscillating disk, let d9 r —r]?(ify foe the velocity of any annulus, and G the moment of Ctu the whole force of the fluid on the disk. Then o / ix/c/o \ i r I -r— dr; o \«#/o an(j f™£\ wiH be got from (188) by substituting rF(t) for /(£). We find thus Q- = — ^Tr// . pa41 Ff (t — £,) -77............(189). Jo v^i If we suppose the angular velocity of the disk to be expressed by A sin nt, where A is constant, we must put F(t) = A sin nt in (189), and we should then get after integration the same expression for Q as was obtained in Art. 8 by a much simpler process. Suppose, however, that previously to the epoch from which t is measured the disk was at rest, and that the subsequent angular velocity is expressed by At sin nt, where At is a slowly varying function of t. Then F(t) = 0 when t < 0, F(t) = At sin nt when t > 0. On substituting in (189) we get At-t, cosn(t — tfj) ~,* ......(190). Now treating At as a slowly varying parameter, we get from a formula given by Mr Airy, and obtained by the method of the variation of parameters, J A n .....................(191), where / denotes the moment of inertia. In the expression for G we may replace At_ti under the integral sign by At outside it, because At is supposed to vary so slowly that At-tl does not much differ from At while ^ is small enough to render the integral of